FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3042768110> ?p ?o ?g. }

• W3042768110 abstract "Recognizing an object's category and pose lies at the heart of visual understanding. Recent works suggest that deep neural networks (DNNs) often fail to generalize to category-pose combinations not seen during training. However, it is unclear when and how such generalization may be possible. Does the number of combinations seen during training impact generalization? Is it better to learn category and pose in separate networks, or in a single shared network? Furthermore, what are the neural mechanisms that drive the network's generalization? In this paper, we answer these questions by analyzing state-of-the-art DNNs trained to recognize both object category and pose (position, scale, and 3D viewpoint) with quantitative control over the number of category-pose combinations seen during training. We also investigate the emergence of two types of specialized neurons that can explain generalization to unseen combinations---neurons selective to category and invariant to pose, and vice versa. We perform experiments on MNIST extended with position or scale, the iLab dataset with vehicles at different viewpoints, and a challenging new dataset for car model recognition and viewpoint estimation that we introduce in this paper, the Biased-Cars dataset. Our results demonstrate that as the number of combinations seen during training increases, networks generalize better to unseen category-pose combinations, facilitated by an increase in the selectivity and invariance of individual neurons. We find that learning category and pose in separate networks compared to a shared one leads to an increase in such selectivity and invariance, as separate networks are not forced to preserve information about both category and pose. This enables separate networks to significantly outperform shared ones at predicting unseen category-pose combinations." @default.
• W3042768110 created "2020-07-23" @default.
• W3042768110 creator A5020830959 @default.
• W3042768110 creator A5024274700 @default.
• W3042768110 creator A5040427850 @default.
• W3042768110 creator A5042578984 @default.
• W3042768110 creator A5043151044 @default.
• W3042768110 creator A5051000016 @default.
• W3042768110 creator A5055261310 @default.
• W3042768110 creator A5070176712 @default.
• W3042768110 creator A5075155488 @default.
• W3042768110 date "2020-07-17" @default.
• W3042768110 modified "2023-09-27" @default.
• W3042768110 title "On the Capability of Neural Networks to Generalize to Unseen Category-Pose Combinations" @default.
• W3042768110 cites W1564611446 @default.
• W3042768110 cites W1578066333 @default.
• W3042768110 cites W1849277567 @default.
• W3042768110 cites W1947481528 @default.
• W3042768110 cites W1990037101 @default.
• W3042768110 cites W1994488211 @default.
• W3042768110 cites W2068777106 @default.
• W3042768110 cites W2105728138 @default.
• W3042768110 cites W2112796928 @default.
• W3042768110 cites W2136123750 @default.
• W3042768110 cites W2194775991 @default.
• W3042768110 cites W2397830550 @default.
• W3042768110 cites W2471821491 @default.
• W3042768110 cites W2560645892 @default.
• W3042768110 cites W2563100679 @default.
• W3042768110 cites W2575842049 @default.
• W3042768110 cites W2607768201 @default.
• W3042768110 cites W2785648721 @default.
• W3042768110 cites W2893064646 @default.
• W3042768110 cites W2913340405 @default.
• W3042768110 cites W2925348834 @default.
• W3042768110 cites W2945033152 @default.
• W3042768110 cites W2949444519 @default.
• W3042768110 cites W2950212750 @default.
• W3042768110 cites W2950812781 @default.
• W3042768110 cites W2962851944 @default.
• W3042768110 cites W2962997291 @default.
• W3042768110 cites W2963177347 @default.
• W3042768110 cites W2963224792 @default.
• W3042768110 cites W2963446712 @default.
• W3042768110 cites W2963712450 @default.
• W3042768110 cites W2963996492 @default.
• W3042768110 cites W2970692043 @default.
• W3042768110 cites W3001083904 @default.
• W3042768110 hasPublicationYear "2020" @default.
• W3042768110 type Work @default.
• W3042768110 sameAs 3042768110 @default.
• W3042768110 citedByCount "3" @default.
• W3042768110 countsByYear W30427681102021 @default.
• W3042768110 crossrefType "posted-content" @default.
• W3042768110 hasAuthorship W3042768110A5020830959 @default.
• W3042768110 hasAuthorship W3042768110A5024274700 @default.
• W3042768110 hasAuthorship W3042768110A5040427850 @default.
• W3042768110 hasAuthorship W3042768110A5042578984 @default.
• W3042768110 hasAuthorship W3042768110A5043151044 @default.
• W3042768110 hasAuthorship W3042768110A5051000016 @default.
• W3042768110 hasAuthorship W3042768110A5055261310 @default.
• W3042768110 hasAuthorship W3042768110A5070176712 @default.
• W3042768110 hasAuthorship W3042768110A5075155488 @default.
• W3042768110 hasConcept C119857082 @default.
• W3042768110 hasConcept C134306372 @default.
• W3042768110 hasConcept C142362112 @default.
• W3042768110 hasConcept C153349607 @default.
• W3042768110 hasConcept C154945302 @default.
• W3042768110 hasConcept C177148314 @default.
• W3042768110 hasConcept C190470478 @default.
• W3042768110 hasConcept C190502265 @default.
• W3042768110 hasConcept C2776035091 @default.
• W3042768110 hasConcept C2781238097 @default.
• W3042768110 hasConcept C2984842247 @default.
• W3042768110 hasConcept C33923547 @default.
• W3042768110 hasConcept C37914503 @default.
• W3042768110 hasConcept C41008148 @default.
• W3042768110 hasConcept C50644808 @default.
• W3042768110 hasConcept C52102323 @default.
• W3042768110 hasConcept C64876066 @default.
• W3042768110 hasConceptScore W3042768110C119857082 @default.
• W3042768110 hasConceptScore W3042768110C134306372 @default.
• W3042768110 hasConceptScore W3042768110C142362112 @default.
• W3042768110 hasConceptScore W3042768110C153349607 @default.
• W3042768110 hasConceptScore W3042768110C154945302 @default.
• W3042768110 hasConceptScore W3042768110C177148314 @default.
• W3042768110 hasConceptScore W3042768110C190470478 @default.
• W3042768110 hasConceptScore W3042768110C190502265 @default.
• W3042768110 hasConceptScore W3042768110C2776035091 @default.
• W3042768110 hasConceptScore W3042768110C2781238097 @default.
• W3042768110 hasConceptScore W3042768110C2984842247 @default.
• W3042768110 hasConceptScore W3042768110C33923547 @default.
• W3042768110 hasConceptScore W3042768110C37914503 @default.
• W3042768110 hasConceptScore W3042768110C41008148 @default.
• W3042768110 hasConceptScore W3042768110C50644808 @default.
• W3042768110 hasConceptScore W3042768110C52102323 @default.
• W3042768110 hasConceptScore W3042768110C64876066 @default.
• W3042768110 hasLocation W30427681101 @default.
• W3042768110 hasOpenAccess W3042768110 @default.
• W3042768110 hasPrimaryLocation W30427681101 @default.

• next